> Yes, that is benign. FLAC is good at compression music, but not so > good at compressing random noise which is used in a couple of the tests.
Indeed, FLAC uses models to predict signal and then encode the residuals which are, hopefully, rather small and require on average fewer bits. For perfectly random noise any model except "verbatim" will yield potentially large residuals, as large as the signal itself, that's why very little or no compression is attained. However, headers might increase slightly the signal length. Regards, Federico _______________________________________________ flac-dev mailing list flac-dev@xiph.org http://lists.xiph.org/mailman/listinfo/flac-dev
